Classification of acoustic emission based partial discharge in oil pressboard insulation system using Fractal Features

Abstract

Oil pressboard insulation used in transformer deteriorates due to partial discharge (PD). This paper reports experimental results and analysis for classification of PDs using acoustic emission (AE) signal of laboratory simulated PDs in oil pressboard insulation system using three different electrode systems. AE signal produced by PD are detected by sensors mounted on the experimental tank surface, stored on an oscilloscope and fed to computer for further analysis. The measured AE signals are analyzed using Fractal Features. A variety of algorithms are available for computation of Fractal Dimension. In this paper, Box counting and Higuchi's algorithm for the determination of fractal dimension, Lacunarity and Approximate Entropy are used for the extraction of fractal features from the time domain PD AE signals. There are significant overlaps of few features for different types of PDs. But few features are distinct for different types of PDs. These features are used for the classification PDs.